If I have the ability to deal infinite damage incrementally, and my opponent can gain infinite life incrementally, can I kill my opponent?

My combo:
Skyshroud Ranger, Tunneling Geopede, Retreat to Coralhelm, and a Boros Garrison in hand.

My opponent's combo:
Anfenza, Kin-Tree Spirit, Kitchen Finks, Viscera Seer

As I understand the rules, this is a segmented loop, and because it needs to be my turn for me to deal damage, my opponent can always stay ahead of me, and I need to eventually stop trying.

However, if I replace Skyshroud Ranger with Sakura-Tribe Scout, I can't kill my opponent on my turn (if my opponent is careful, otherwise I could response to a persist trigger), but I could kill my opponent on their upkeep.

I think you're wrong on the second point. they can do their entire combo while an iteration of yours is on the stack, so in response to any attempt to deal 1 damage, they can gain infinite life. it's not a fragmented loop, they just combo out forever any time you try to start.

The most useful answer I can give to this kind of questions is "If this scenario ever happens to you, pinch yourself, because chances are you are not living real life!"


The question here is "Can you come up with a shortcut that ends up with your opponent dead? And if so, will your opponent be willing/forced to accept it?"

You may say: "Hi! This is my shortcut. I activate Skyshroud Ranger, put Boros Garrison onto the battlefield, return it to my hand, untap Skyshroud Ranger, Geopede deals 1 damage to you. Repeat a million times. You're dead and I'm so cool ".

And your opponent will answer: "Nope, sorry but I'm cooler than you. In response to each one of your activations of Skyshroud Ranger I sacrifice Kitchen Finks, it returns to the battlefield, I gain two life, Anafenza triggers, I put a +1/+1 counter on Kitchen Finks, the counters annihilate each other. Repeat a billion times. So I end up gaining two billion life for each 1 damage you deal to me, and you can't stop me "

And you will reply: "OK. I can't stop you "


If you try to combo on your opponent's turn, same as above, your opponent can turn your shortcut into one where she gains more life than she loses.

However, if you opponent wants to start her combo first, you may kill her in response, no matter whose turn it is.

So the trick here is "Just wait until your opponent pulls the trigger first!".

Yes, but as soon as they say "I gain a billion life", I tell them, ok, I begin again to deal you 2 billion. And then they respond, and then I try again. Does the game end in a draw?

right, but each loop is separate. 716.3 (where fragmented loops are discussed) only talks about scenarios where the same game state is reached multiple times. in this case that's not what's happening: each time you iterate the process you wind up with a different life total, so it's not the same thing.

So it is a draw? Because I thought of that interpretation, but a draw occurring when both players are free to stop (and especially stop via inaction) is odd.

Basically if you try to combo first, your opponent may "fix" your shortcut so that on each iteration she ends up gaining more life than she loses. Now you can state how many times you want to repeat the entire loop (maybe just one is enough, maybe not).

In the end, this is equivalent to "you choose a number for your loop, then your opponent chooses a (probably higher) number for her loop, then the result is processed and both loops must stop", so if you try it first you are screwed.

But this scenarios won't happen in real life. Real life is happier. That's why the guys who make the rules don't care too much about them.

This is exact scenario happened to me. Only on MTGO, so it was a battle of the clocks.

Judge Chat seems to agree that this qualifies as a segmented loop and that "Game State" should be interpreted loosely. I'm curious as to what Zammm's thoughts might be though.

My thoughts are that his is exactly the kind of corner case that people love to spend way, way more effort on than it's actually worth. In short:

Yes, it does happen occasionally. But not anywhere near often enough to be worth the sort of time and mental effort people devote to it.

Though they're technically covered in the CR, in practice loops are way more of a judging issue than a game rules one, in much the same way that "Handling Illegal Actions" is, despite having a dedicated portion of the CR all to itself. Handling loops is way more about dealing with player behavior than it is dealing with the outcomes of actions, and as such, their CR coverage is relatively perfunctory--if you really start trying to deliberately poke holes in the "loop rules", you'll find them to be woefully weird and inadequate. (For example, they just outright cannot handle a loop that occurs over the course of multiple turns.)

I suspect that if it ever ended up actually mattering--it's not going to, but if it did--the relevant people would come up with a much more sensible answer than whose turn it is determining whether or not somebody can win the game--the "nonactive player wins" bit is obviously intended for cases more like the example it provides in the rules than it is for something like this.

I'm not well-versed in the tournament rules, but I'd like to think that in paper Magic, Dr_Demento would get a slow play warning for the first scenario. It doesn't make sense to let you continuously do something that you know won't progress the game. I know technically the game could end if your opponent decided to give up and lose, but they're not the one forcing the game into an endless loop. Like, if you found a way to infinitely tap and untap your opponent's Fallowsage, the game shouldn't end in a draw just because your opponent refuses to deck themself.

That's just my opinion, though, and of course MTGO is another matter entirely.

If only Murderous Redcap couldn't deal damage to creatures, this issue would find resolution

a) just use gravelgill axeshark and purphoros, god of the forge, works basically the same way, and b) I still don't think it matters because either combo can go off infinite times in response to the other.

I think Jim's basically right. you go "I declare this loop that will deal 1 damage a hundred times" and they go "ok modify that loop so that every time you deal 1 damage I gain a thousand life." then you can do the loop if you want, but at the end they're at a higher life total than they started. you can then initiate again, they modify again, and so on. each time you initiate they wind up gaining life, so at some point you refusing to move on becomes slow play because no iteration of the loop is strategically valuable to you. you're technically affecting the game by raising their life total, you're not just twiddling, but unless you can show that that's somehow valuable to you it's still basically slow play.

Oh, this thing actually does happen, but if it legitimately happened regularly at GPs then there would be incentive to come up with a definitive answer. Currently there is an incrementive instant speed life gain combo and and incrementive instant speed damage combo. However, it works itself out because the damage combo can just kill all the life gain combo's creatures and then finish them off. If Redcap couldn't kill creatures (and was still played), then this would be a real problem.

Also, you interpretation of segmented loops makes them essentially worthless, because almost all instances of them change the game state (although in Standard, Sylvan Advocate and Wandering Fumerole have potential). If I have a way to make infinite mana and a Bellows Lizard and I attack into my opponent's Drifting Shade when my opponent can also generate mana, they are forced to trade, but if it is a Nightwing Shade they can save it?